Both nodes operates on ExampleSets in RapidMiner, and most common cells in KNIME (Int, Double, and String). In both nodes, the data available is limited by the memory of the JVM, they do not try to put part of the data to the disks to perform the tasks.

RapidMiner node

  • The range and other domain related information are not propagated to RapidMiner
    • The information not present cannot be used in correctness checking
  • The output do not contain role information
  • The properties are transferred to and from this node
  • The RapidMiner preferences cannot be set currently
    • so behind a proxy, the network connections from RapidMiner might fail
    • parallelism settings
    • localisation
    • tools
    • logging
    • GUI preferences
    • database drivers, connections
    • manage extensions
  • Cannot load processes from the RapidMiner repositories yet
  • After changing the main RapidMiner location we have to restart KNIME to reinitialize the RapidMiner instances
  • Sometimes the previously computed column specification differs from the result's column specification, which causes a confusing message, but it works well with the result data

RapidMiner Viewer node

  • The Highliting information is not shown on the diagrams, cannot highlite from this node
  • The colour, shape transformations are not used